3D shape constraint for facial feature localization using probabilistic-like output

Longbin Chen, Lei Zhang, Hongjiang Zhang, Mohamed Abdel-Mottaleb

Research output: Chapter in Book/Report/Conference proceedingConference contribution

23 Citations (Scopus)

Abstract

This paper presents a method to automatically locate facial feature points under large variations in pose, illumination and facial expressions. First we propose a method to calculate probabilistic-like output for each pixel of image. This probabilistic-like output describes the possibility of the pixel to be the center of specified object. A Gaussian Mixture Model is used to approximate the distribution of probabilistic-like output. The centers of these Gaussians are assigned with a probabilistic-like measure and they are considered as candidate feature points. There might be one or more candidate feature points in each facial region. A 3D model of facial feature points is built to enforce constraints on the localization results of feature points. Compared with Active Shape Model (ASM) and its variant methods, our method could accommodate larger variations in pose, lighting and face expressions. Moreover, it is less sensitive to initialization errors, accurate, and fast. It takes a computer with P4 CPU about 10ms to locate the five feature points (two eye centers, two mouth corners and nose tip). The feature localization accuracy is comparable with the accuracy of manually labeled features and it is robust to noise (glasses, beards). Experiments on FERET gallery and PIE are reported in this paper as well.

Original languageEnglish
Title of host publicationProceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition
Pages302-307
Number of pages6
StatePublished - Sep 24 2004
EventProceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition FGR 2004 - Seoul, Korea, Republic of
Duration: May 17 2004May 19 2004

Other

OtherProceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition FGR 2004
CountryKorea, Republic of
CitySeoul
Period5/17/045/19/04

Fingerprint

Lighting
Pixels
Program processors
Glass
Experiments

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Chen, L., Zhang, L., Zhang, H., & Abdel-Mottaleb, M. (2004). 3D shape constraint for facial feature localization using probabilistic-like output. In Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition (pp. 302-307)

3D shape constraint for facial feature localization using probabilistic-like output. / Chen, Longbin; Zhang, Lei; Zhang, Hongjiang; Abdel-Mottaleb, Mohamed.

Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition. 2004. p. 302-307.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Chen, L, Zhang, L, Zhang, H & Abdel-Mottaleb, M 2004, 3D shape constraint for facial feature localization using probabilistic-like output. in Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition. pp. 302-307, Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition FGR 2004, Seoul, Korea, Republic of, 5/17/04.
Chen L, Zhang L, Zhang H, Abdel-Mottaleb M. 3D shape constraint for facial feature localization using probabilistic-like output. In Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition. 2004. p. 302-307
Chen, Longbin ; Zhang, Lei ; Zhang, Hongjiang ; Abdel-Mottaleb, Mohamed. / 3D shape constraint for facial feature localization using probabilistic-like output. Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition. 2004. pp. 302-307
@inproceedings{98fae8bef83043089570768d1a10a56d,
title = "3D shape constraint for facial feature localization using probabilistic-like output",
abstract = "This paper presents a method to automatically locate facial feature points under large variations in pose, illumination and facial expressions. First we propose a method to calculate probabilistic-like output for each pixel of image. This probabilistic-like output describes the possibility of the pixel to be the center of specified object. A Gaussian Mixture Model is used to approximate the distribution of probabilistic-like output. The centers of these Gaussians are assigned with a probabilistic-like measure and they are considered as candidate feature points. There might be one or more candidate feature points in each facial region. A 3D model of facial feature points is built to enforce constraints on the localization results of feature points. Compared with Active Shape Model (ASM) and its variant methods, our method could accommodate larger variations in pose, lighting and face expressions. Moreover, it is less sensitive to initialization errors, accurate, and fast. It takes a computer with P4 CPU about 10ms to locate the five feature points (two eye centers, two mouth corners and nose tip). The feature localization accuracy is comparable with the accuracy of manually labeled features and it is robust to noise (glasses, beards). Experiments on FERET gallery and PIE are reported in this paper as well.",
author = "Longbin Chen and Lei Zhang and Hongjiang Zhang and Mohamed Abdel-Mottaleb",
year = "2004",
month = "9",
day = "24",
language = "English",
isbn = "0769521223",
pages = "302--307",
booktitle = "Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition",

}

TY - GEN

T1 - 3D shape constraint for facial feature localization using probabilistic-like output

AU - Chen, Longbin

AU - Zhang, Lei

AU - Zhang, Hongjiang

AU - Abdel-Mottaleb, Mohamed

PY - 2004/9/24

Y1 - 2004/9/24

N2 - This paper presents a method to automatically locate facial feature points under large variations in pose, illumination and facial expressions. First we propose a method to calculate probabilistic-like output for each pixel of image. This probabilistic-like output describes the possibility of the pixel to be the center of specified object. A Gaussian Mixture Model is used to approximate the distribution of probabilistic-like output. The centers of these Gaussians are assigned with a probabilistic-like measure and they are considered as candidate feature points. There might be one or more candidate feature points in each facial region. A 3D model of facial feature points is built to enforce constraints on the localization results of feature points. Compared with Active Shape Model (ASM) and its variant methods, our method could accommodate larger variations in pose, lighting and face expressions. Moreover, it is less sensitive to initialization errors, accurate, and fast. It takes a computer with P4 CPU about 10ms to locate the five feature points (two eye centers, two mouth corners and nose tip). The feature localization accuracy is comparable with the accuracy of manually labeled features and it is robust to noise (glasses, beards). Experiments on FERET gallery and PIE are reported in this paper as well.

AB - This paper presents a method to automatically locate facial feature points under large variations in pose, illumination and facial expressions. First we propose a method to calculate probabilistic-like output for each pixel of image. This probabilistic-like output describes the possibility of the pixel to be the center of specified object. A Gaussian Mixture Model is used to approximate the distribution of probabilistic-like output. The centers of these Gaussians are assigned with a probabilistic-like measure and they are considered as candidate feature points. There might be one or more candidate feature points in each facial region. A 3D model of facial feature points is built to enforce constraints on the localization results of feature points. Compared with Active Shape Model (ASM) and its variant methods, our method could accommodate larger variations in pose, lighting and face expressions. Moreover, it is less sensitive to initialization errors, accurate, and fast. It takes a computer with P4 CPU about 10ms to locate the five feature points (two eye centers, two mouth corners and nose tip). The feature localization accuracy is comparable with the accuracy of manually labeled features and it is robust to noise (glasses, beards). Experiments on FERET gallery and PIE are reported in this paper as well.

UR - http://www.scopus.com/inward/record.url?scp=4544340559&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=4544340559&partnerID=8YFLogxK

M3 - Conference contribution

SN - 0769521223

SP - 302

EP - 307

BT - Proceedings - Sixth IEEE International Conference on Automatic Face and Gesture Recognition

ER -